{"id":"https://openalex.org/W4408354070","doi":"https://doi.org/10.1109/icassp49660.2025.10890743","title":"Speech-N-LlaMA: Improving Speech LLMs with Multi-Pass Training","display_name":"Speech-N-LlaMA: Improving Speech LLMs with Multi-Pass Training","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408354070","doi":"https://doi.org/10.1109/icassp49660.2025.10890743"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10890743","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10890743","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5055125973","display_name":"Amit Kumar Singh Yadav","orcid":"https://orcid.org/0000-0001-6464-7688"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Amit Kumar Singh Yadav","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5048538280","display_name":"Gil Keren","orcid":"https://orcid.org/0000-0002-5153-3494"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gil Keren","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5004777817","display_name":"Desh Raj","orcid":"https://orcid.org/0000-0002-5038-9400"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Desh Raj","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026526428","display_name":"Wei Zhou","orcid":"https://orcid.org/0000-0003-3622-3970"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei Zhou","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5113970008","display_name":"Junteng Jia","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Junteng Jia","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100343462","display_name":"Ke Li","orcid":"https://orcid.org/0000-0001-9206-0892"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ke Li","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020171946","display_name":"Ying Xu","orcid":"https://orcid.org/0000-0002-7856-3343"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ying Xu","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101579713","display_name":"Chunyang Wu","orcid":"https://orcid.org/0000-0002-0269-3555"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chunyang Wu","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5074237839","display_name":"Jay Mahadeokar","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jay Mahadeokar","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5066166549","display_name":"Ozlem Kalinli","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ozlem Kalinli","raw_affiliation_strings":["Meta,USA"],"affiliations":[{"raw_affiliation_string":"Meta,USA","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":10,"corresponding_author_ids":["https://openalex.org/A5055125973"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.01907413,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9990000128746033,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.9990000128746033,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9810000061988831,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9556000232696533,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.67787104845047},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.623536229133606},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.6163257360458374},{"id":"https://openalex.org/keywords/speech-synthesis","display_name":"Speech synthesis","score":0.5290723443031311},{"id":"https://openalex.org/keywords/audiology","display_name":"Audiology","score":0.3256981670856476},{"id":"https://openalex.org/keywords/medicine","display_name":"Medicine","score":0.13614752888679504}],"concepts":[{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.67787104845047},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.623536229133606},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.6163257360458374},{"id":"https://openalex.org/C14999030","wikidata":"https://www.wikidata.org/wiki/Q16346","display_name":"Speech synthesis","level":2,"score":0.5290723443031311},{"id":"https://openalex.org/C548259974","wikidata":"https://www.wikidata.org/wiki/Q569965","display_name":"Audiology","level":1,"score":0.3256981670856476},{"id":"https://openalex.org/C71924100","wikidata":"https://www.wikidata.org/wiki/Q11190","display_name":"Medicine","level":0,"score":0.13614752888679504},{"id":"https://openalex.org/C153294291","wikidata":"https://www.wikidata.org/wiki/Q25261","display_name":"Meteorology","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49660.2025.10890743","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10890743","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4","score":0.44999998807907104}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":29,"referenced_works":["https://openalex.org/W1494198834","https://openalex.org/W2103869314","https://openalex.org/W2127141656","https://openalex.org/W2799473636","https://openalex.org/W2952649152","https://openalex.org/W2962824709","https://openalex.org/W2963414781","https://openalex.org/W2963747784","https://openalex.org/W3011339933","https://openalex.org/W3095410713","https://openalex.org/W3097777922","https://openalex.org/W3197267207","https://openalex.org/W4293363567","https://openalex.org/W4372266855","https://openalex.org/W4389520395","https://openalex.org/W4391021574","https://openalex.org/W4391021666","https://openalex.org/W4392903956","https://openalex.org/W4392931626","https://openalex.org/W6621543089","https://openalex.org/W6637373629","https://openalex.org/W6675365184","https://openalex.org/W6679436768","https://openalex.org/W6685322675","https://openalex.org/W6796581206","https://openalex.org/W6847363464","https://openalex.org/W6850625674","https://openalex.org/W6853998256","https://openalex.org/W6854445763"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W230091440","https://openalex.org/W2390279801","https://openalex.org/W2233261550","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2810751659"],"abstract_inverted_index":{"Speech":[0,35,68,103,144],"LLMs":[1,69],"use":[2],"speech":[3,21],"embeddings":[4],"as":[5,54],"the":[6,20,41,66,75,92,100],"prompt":[7],"to":[8,40,65,98,117,132,142],"a":[9,29,47],"Large":[10],"Language":[11],"Model":[12],"(LLM)":[13],"and":[14,51,90,110,129],"generate":[15],"human":[16],"readable":[17],"text":[18],"for":[19,33],"signal":[22],"in":[23,115,136],"an":[24,107],"autoregressive":[25],"manner.":[26],"Teacher-forcing":[27],"is":[28,38],"common":[30],"approach":[31],"used":[32,43],"training":[34,50],"LLMs,":[36],"which":[37],"dissimilar":[39],"procedure":[42],"during":[44,77,152],"inference,":[45],"creating":[46],"gap":[48],"between":[49],"inference":[52],"known":[53],"exposure":[55,59,88],"bias.":[56],"To":[57],"mitigate":[58],"bias,":[60],"we":[61],"propose":[62,106],"Speech-N-LlaMA.":[63],"Contrary":[64],"existing":[67],"that":[70],"have":[71],"single":[72],"pass":[73],"through":[74],"LLM":[76,97],"training,":[78],"Speech-N-LlaMA":[79,86,116],"incorporates":[80],"multi-pass":[81],"training.":[82],"Through":[83],"multiple":[84],"passes,":[85],"mitigates":[87],"bias":[89],"uses":[91],"error":[93],"correction":[94],"capability":[95],"of":[96,102],"improve":[99],"performance":[101],"LLMs.":[104],"We":[105,120],"N-pass":[108],"loss":[109],"utterance":[111],"level":[112],"temperature":[113],"sampling":[114],"achieve":[118],"this.":[119],"evaluate":[121],"four":[122],"different":[123],"model":[124],"sizes":[125],"on":[126],"three":[127],"benchmarks,":[128],"show":[130],"up":[131],"18%":[133],"relative":[134],"improvement":[135],"Word":[137],"Error":[138],"Rate":[139],"(WER)":[140],"compared":[141],"baseline":[143],"LLM,":[145],"while":[146],"not":[147],"incurring":[148],"any":[149],"additional":[150],"compute":[151],"inference.":[153]},"counts_by_year":[],"updated_date":"2025-12-19T19:40:27.379048","created_date":"2025-10-10T00:00:00"}